363
Acc. Chem. Res. 1990,23, 363-369 insulating substrates by measurement of the feedback current really represent two limiting cases where the heterogeneous electron transfer (et) to tip-generated species is either very fast or very slow. However, measurements at finite et rates at the substrate are also possible and lead to steady-state iT vs d curves that span the region between the limiting cases. Similarly it should be possible to measure the porosity of a nonconductive film coating a conductive substrate. If the tip radius, a, is small compared to the pore size, the actual pore distribution can be determined. Even for pores that are small compared to a, information about the porosity should be obtainable from the iT vs d curve. SECM studies of polymer films or other types of modified electrode surfaces have been described earlier. It should also be possible to modify the tip electrode, e.g., with an organized monolayer or polymer film, and study et reactions between it and a suitable substrate. Such studies would be related to those carried out in the ultrathin-layer cellz7but, with proper tip electrode construction, may be somewhat easier to execute. While most SECM studies involved steady-state currents and studies of substrate properties, measurements of current transients should also be informative, e.g., in determination of diffusion coefficients and rate (27) Fan,
F.-R. F.; Bard, A. J. J. Am. Chem. SOC.1987, 109, 6262.
constants. Measurement of steady-state and transient currents might also be used to characterize homogeneous reactions of the tip-generated species that occur in the solution gap between the tip and substrate, in a manner analogous to measurements made with the rotating ring-disk electrode.10*28These applications await the development of suitable models, probably digital simulation ones. Finally, many applications to the characterization and modification of insulating and conducting substrates are possible. The ultimate resolution possible with the SECM depends upon our ability to construct scanning tips of a small size of the proper shape. While it is unlikely that the resolution will ever approach that of the STM, improvements to bring SECM to the levels of several tens of nanometers, by reduction in tip size and by employing deconvolution or tomographic-type techniques, should be possible. W e gratefully acknowledge the support of this research by the National Science Foundation (CHE 8901450), the Robert A. Welch Foundation, and the Texas Advanced Research Program. D.M. acknowledges the support of a Chaim Weizmann Fellowship. W e also acknowledge the important contributions of Drs. J. Kwak and F. R. Fan to the development of the SECM at The University of Texas at Austin. (28) Albery, W. J.; Hitchman, M. L. Ring-Disc Electrodes;Clarendon: Oxford, 1971.
Structure and Mechanism of Action of a Multifunctional Enzyme: Fatty Acid Synthase SOO-IK CHANG Center for Biochemical and Biophysical Sciences and Medicine, Harvard Medical School, Boston, Massachusetts 021 15
GORDONG. HAMMES* Department of Chemistry, University of California, Santa Barbara, Santa Barbara, California 93106 Received April 3, 1990 (Revised Manuscript Received July 30, 1990)
The synthesis of fatty acids is a ubiquitous process in nature. However, a variety of different enzymes catalyze the synthesis, and their structures display considerable variation. In this account, we will be concerned with the fatty acid synthase from chicken liver which primarily produces palmitic acid. In most procaryotes, the synthesis of palmitic acid is carried out Soo-Ik Chang received his B.S. and M.S. degrees (with Prof. TongSeek Chair) from Korea University, Seoul, Korea. He completed his m.D. degree in 1989 at Cornell University under the guidance of Prof. Gordon G. Hammes and is currently a postdoctoral research fellow with Prof. Bert L. Vallee at Harvard Medical School. His research interests include studies of the structure and function of chicken liver fatty acid synthase and the interaction of angiogenin with cell surface receptors. He was the recipient of a fellowship of the Ministry of Education of Korea (1983-1987). Gordon G. Hammes is a Professor of Chemistry and Biological Sciences and Vice Chancellor for Academic Affalrs at the University of California, Santa Barbara. He was previously a member of the faculties at Massachusetts Institute of Technology and Cornell University. His research interests have centered around the kinetics and mechanisms of enzymatic reactions, including studies of very fast reactions, physical chemical characterization of proteins, membrane-bound enzymes, and multienzyme complexes.
0001-4842/90/0123-0363$02.50/0
by seven different and separable monofunctional enzymes and an acyl carrier protein (cf. refs 1 and 2). However, in eucaryotes a multifunctional enzyme is involved that contains the seven enzyme activities. In yeast and other fungi, a multifunctional enzyme of two different types of polypeptides is found (aSps),3whereas in animals, the enzyme consists of two identical polypeptides (az).1i2The multifunctional enzymes in animals catalyze a reaction sequence that leads to palmitic acid according to the overall reaction acetyl-coA + 7malonyl-CoA + 14NADPH + 14H' palmitic acid + 8CoA + 14NADP+ + 6Hz0 + 7 c o z (1)
-
(1)Wakil, S. J.; Stoops, J. K.;Joshi, V. C. Annu. Rev. Biochem. 1983, 52,537. (2) Wakil, S.J. Biochemistry 1989,28, 4523. (3) Schweizer, M.Multidomain Proteins-Structure and Evolution; Hardie, D. G., Coggins, J. R.,Ed.; Elsevier: New York, 1986; p 195.
0 1990 American Chemical Society
364 Acc. Chem. Res.,
Vol.23, No.11, 1990
Chang and Hammes 0
I
TE
SYNTHASE
loo0
500
A/MT
KS
SH
I500 I
I
DH
2438
Zoo0
ER
KR
TE
ACP
U;
I
I
OH
I
2438
2d00
lh
iI idoo
sbo
b
I
TE
DH
SYNTHASE
Figure 1. Schematic representation of the fatty acid synthase structure. The two polypeptide chains and the binding sites that have been labeled with fluorescent/absorbance probes are shown, as well as some of the measured distances. The dashed line separates the two independent catalytic centers, each composed of two different polypeptide chains. DH is dehydratase; ER is enoyl reductase; KR is @-ketoacylreductase; T E is thioesterase; PMP is pyridoxamine phosphate (site I); and -SH is 4'-phosphopantetheine. The essential cysteine (-SH) and serine (-OH) are shown.1°
Multifunctional enzymes have several potential physiological advantages, for example, enhancement of the intrinsic catalytic activity of individual enzymes, sequestering of reaction intermediates, and regulation of metabolic fluxes. Palmitic acid is generated by initiating the fatty acid chain with the acetyl group from acetyl-coA and adding two carbons at a time from seven malonyl-CoA molecules. The growing fatty acid chain is covalently bound to the enzyme: during the catalytic reactions, it is attached through a thioester linkage to the sulfhydryl group of 4'-phosphopantethoic acid which is covalently linked to a serine residue of the protein. The specific sequence of reactions is as follows: acetyl-coA + E + acetyl-E + CoA (2) malonyl-CoA + acetyl-E + acetyl-E-malonyl + CoA (3) acetyl-E-malonyl T- CH3C(0)CH2C(O)-E+ C02 (4) CH,C(O)CH&(O)-E + NADPH + H+ ----L CH&H(OH)CH&(O)-E + NADP' (5) CH,CH(OH)CH,C(O)-E + CH,CH=CHC(O)-E + H20 (6) CH&H=CHC(O)-E + NADPH + H+ CH&H2CH&(O)-E + NADP' (7) The sequence of enzyme activities is as follows: acetyl transacylase, malonyl transacylase, 6-ketoacyl synthase, P-ketoacyl reductase, dehydratase, and enoyl reductase. In the chicken liver enzyme, the acetyl and malonyl transacylase activities are combined in a single enzyme! When the fatty acid chain of the intermediate is saturated (eq 7), a malonyl residue is transferred to the enzyme (eq 3) and two more carbons are added. This cycling continues until a palmitoyl thioester is formed, which is cleaved by a palmitoyl thioesterase to give the free acid. Structure of Fatty Acid Synthase Chicken liver fatty acid synthase has a molecular weight of about 534000 and consists of two identical (4) Chang,
%I.; Hammes, G. G. Biochemistry 1988, 27,4753.
Figure 2. Functional map of chicken liver fatty acid synthase. Boxes are regions of great homology with the rat enzyme, whereas the lines identify less homologous regions and are often connections between structural domains. The essential cysteine (-SH), serine (-OH), 4'-phosphopantetheine (-SH),and NADPH binding sites are shown. KS is @-ketoacylsynthase; A/MT is acetyl/malonyl transacylase; DH is dehydratase; ER is enoyl reductase; KR is @-ketoacylreductase; ACP is acyl carrier protein; and T E is thioesterase. The numbering of the amino acids also is given.% The rat polypeptide has 74 additional amino acids on the N-terminus.20 If these amino acids also are present on the chicken liver polypeptide, the numbering of amino acids will be changed.
polypeptide chains (cf. refs 1,2,and 5). These polypeptides are arranged head to tail and combine to give two independent catalytic centers: each derived from two different polypeptides. This is illustrated in Figure 1where the two independent catplytic centers are separated by a vertical dashed line. This structure was deduced from many different lines of evidence. For example, if the two polypeptide chains are separated, all of the catalytic activities are intact except for the P-ketoacyl synthase and the P-ketoacyl reductase.ls7 Also, the cysteine on the P-ketoacyl synthase of one polypeptide chain has been cross-linked to the 4'phosphopantetheine on a different polypeptide chain indicating that these with 1,3-dibrom0-2-propanone,~ two sulfhydryls are within 5 A of each other. The cDNA coding for the polypeptide has been sequenced and the corresponding amino acid sequence d e d u ~ e d . ~The , ~ active-site regions of the component enzymes, which are indicated in Figures 1 and 2, were primarily identified from amino acid sequencing of active site labeled peptide~.~JOStarting from the Cterminus, the enzymes are thioesterase, P-ketoacyl reductase, enoyl reductase, dehydratase, acetyl/malonyl transacylase, and P-ketoacyl synthase. The growing fatty acid chain is bound to 4'-phosphopantetheine in a region between the thioesterase and P-ketoacyl reductase; this region is often called the "acyl carrier protein" because an isolatable protein carrying out this function is found in Escherichia coli. This sequence of activities also has been deduced from studies of limited proteolytic digestion of the multifunctional enzyme.lY2J1 The three-dimensional structure of the enzyme remains to be elucidated. However, small-angle neutron scattering and electron microscope studies suggest that (5) Holzer, K. P.; Liu, W.; Hammes, G. G. R o c . Natl. Acad. Sci. U.S.A. 1989,86,4387. (6) Singh, N.; Wakil, S.J.; Stoops, J. K. J.Biol. Chem. 1984,259,3605. (7) Kashem, M. A.; Hammes, G. G. Biochim. Biophys. Acta 1988,956, 39. (8) Stoops, S. J.; Wakil, J. K. J. Biol. Chem. 1981,256, 5128. (9) Yuan, Z.; Liu, W.; Hammes, G. G. Roc. Natl. Acad. Sci. U.S.A. 1988,85,6328. (10) Chang, &-I.; Hammes, G. G. Biochemistry 1989,28, 3781. (11) Tsukamoto, Y.; Wong, H.; Mattick, J. S.; Wakil, S. J. J. Biol. Chem. 1983,258,15312.
Fatty Acid Synthase
Ace. Chem. Res., Vol. 23, No. l l , 1990 365
lypeptides from yeast and animals have evolved from the polypeptides have dimensions of 160 X 146 X 73 A the monofunctional enzymes of lower s p e c i e ~ . 4 J ~ - ~ ~ with three domains, 32, 82, and 46 8, in length.12 In Recently, the sequencing of the full cDNA encoding addition, fluorescence resonance energy transfer and the chicken l i ~ e r ,rat,20 ~ ? ~ and yeast fatty acid synelectron paramagnetic resonance have been used to t h a ~ e ~has l - ~been ~ reported. To obtain further infordetermine the distances between specific sites on the mation about conserved functionally important domains er~zyme.’~J~J~ and evolutionary relationships among the enzymes, a Because of the size of the protein, specific chemical homology analysis of the protein sequences has been labeling is difficult. However, a number of fluorescent carried The entire amino acid sequences of the and ultraviolet-visible absorbing probes that label chicken and rat enzymes are 67% identical. If conspecific sites were developed and their locations deservative substitutions are allowed, 78% of the amino termined by protein sequencing. The sites labeled are acids are matched. A region of low homologies exists as follows: 4’-phosphopantetheine, serine at the active between the functional domains. In particular, a string site of the thioesterase, lysine at the active site of enoyl of 206 amino acids in a region of low homologies is reductase, and a specific lysine (site I) located in a located at amino acid residues 1059-1264 of the chicken domain on the N-terminus half of the polypeptide.1°J3 enzyme. This region connects two functionally different In addition, a variety of analogues of NADPH, includparts of the polypeptide: that is, as shown in Figures ing spin labels, were ~tilized.’~J~ The distances deter1 and 2, in forming a multienzyme reaction center for mined between specific sites and the locations of the fatty acid synthesis, the P-ketoacyl synthase and acesites within the structural model are included in Figure tyl/malonyl transacylase interact with a different po1. The 4’-phosphopantetheine is about 20 A long so lypeptide containing the other enzymes and acyl carrier. that if the growing chain is to reach all of the catalytic The arrangement of the catalytic sites of chicken liver sites, they must lie within a circle 40 A in diameter. fatty acid synthase by homology analysis of the protein Some of the measured distances exceed this dimension; sequence with the rat enzyme is represented in Figure although the uncertainty is sufficient to prevent a firm 2. Boxes and lines are regions of high and low homconclusion, relatively large conformational changes may ologies with the rat enzyme, respectively, as defined in be part of the catalytic cycle. (The uncertainty in the ref 24. The connecting regions between the individual distances due to experimental and theoretical factors enzymes of the rat and chicken fatty acid synthase have is about 10-20%. However, the fluorescence probes are relatively low homologies. This suggests either that rather large so that the precise location of the transition during evolution different joining events occurred for dipole relative to the specific amino acid residue is the two species or that the rat and chicken synthase unknown. This can lead to uncertainties as large as share a common ancestor with a high rate of mutation 10-20 8, in locating the specific sites.) occurring in the joining regions. A higher rate of muThe thioesterase domain is at the C-terminus and can be readily cleaved from the enzyme by prote~lysis.’*~*’~tations in the joining sequences than in the active sites of the enzymes without loss of function would not be If the thioesterase is cleaved from the native enzyme unexpected. with trypsin, it cannot interact with the rest of the A high degree of homology exists between the active enzyme to hydrolyze the long-chain fatty acyl thioester. sites of the chicken and rat enzymes, supporting the The “clipped” enzyme still performs all the catalytic hypothesis that structurally conserved enzymes were reactions necessary to synthesize a long-chain fatty acyl combined during the evolutionary process. A lower thioester. Dynamic fluorescence anisotropy measuredegree of homology exists between the active sites of ments of the enzyme, with the thioesterase active site the chicken and yeast enzymes, suggesting that either serine specifically labeled with a fluorescent probe, inthe yeast fatty acid synthase diverged from the evoludicate a rotational correlation time of the intact thiotionary path of the rat and chicken enzymes at a very esterase of 44 ns,13 This is not very different from the early time during the joining process or the joining rotational correlation time of the trypsin fragment process occurred on a different evolutionary pathway. containing the thioesterase, 17 ns, but is much shorter than that for the entire enzyme, 610 ns.16 This indiMechanism of Action of Fatty Acid Synthase cates that the thioesterase has considerable freedom of motion and therefore is capable of large conformational During the catalytic cycle, the reaction intermediates changes within the multifunctional enzyme during caare covalently attached to three different sites on the talysis. Such a conformational change might be imenzyme: the acetyl and malonyl residues are loaded portant during catalysis since the thioesterase site is relatively far from 4’-phosphopantetheine (Figure 1). (17)Hardie, D. G.; McCarthy, A. D. Multidomain Proteim-
Evolution of Fatty Acid Synthase The evolutionary aspects of fatty acid synthase are of interest in determining how multifunctional enzymes were developed from individual enzymes. The strong similarities between several substrate binding sites from different species suggest that the multifunctional po(12)Stoop, J. K.; Wakil, S.J.; Uberbacher, E. C.; Bunick, G. J. J. Biol. Chem. 1987,262, 10246. (13)Yuan, 2.;Hammes, G. G. J. Biol. Chem. 1986,261, 13643. (14)Chang, &-I.;Hammes, G. G. Biochemistry 1986,264661. (15)Lin, C. Y.;Smith, S. J . Biol. Chem. 1978, 253, 1954. (16)Anderson, V. E.;Hammes, G. G. Biochemistry 1983, 22, 2995.
Structure and Evolution; Hardie, D. G.,and Coggins, J. R., Eda.;Elsevier: New York, 1986;p 229. (18)Poulose,A. J.; Bonsall, R. F.; Kolattukudy, P. E. Arch. Biochem. Biophys. 1984,230, 117. (19)Witkowski, A.; Naggert, J.; Mikkelsen, J.; Smith, S. Eur. J. Biochem. 1987,165,601. (20)A m y , C. M.; Witkowski, A.; Naggert, J.; Williams, B.; Randhawa, Z.: Smith. S. h o c . Natl. Acad. Sci. U.S.A. 1989,86, 3114. ’ (21)Schweizer, M.; Roberta, L. N.; Holtke, H.-J;; Takabayashi, K.; Hollerer, E.; Hoffmann, B.; Muller, G.; Koltig, H.; Schwezer, E. Mol. Gen. Genet. 1986, 203, 479. (22)Chirala, S.S.;Kuziora, M.; Spector, D.; Wakil, S.J. J. Biol. Chem. 1987.262.4231. ,---, (23)Mohamed, A. H.; Chirala, S. S.;Mody, N. H.; Huang, W.-Y.; Wakil, S. J. J. Biol. Chem. 1988, 263, 12315. (24)Chang, %I.; Hammes, G. G. R o c . Natl. Acad. Sci. U.S.A. 1989, 86, 8373.
-
-
Chang and Hammes
366 Ace. Chem. Res., Vol. 23, No. 11, 1990 HS
OH
::
CH3-C-CoA
HS
Acetyl /Malonyl Transacylase
0 OH CH3-C-S
0 HO-C-CH2-C-CoA II II
II
0
CH3-C-S
HS
I. B - K a t o o c y l
OH
I
O-C-CH2-C-OH
1 ;
I
HS
II
0
II 0
II 0
2.D e h y d r a i a s e 3 . E n o y l reductose
HS
OH
z
-
CH3-C-S II
I
1
OH
B-Ketoacyl
0
S-C-CH2-C-OH II 0
I
-
OH
CH3-(CH212-C-S
OH
0 II 0
- HS
OH
Thioesterase
- -
S C (CH2114- C H 3 II
0
____. Synthase
0
__*
S-C-CH2-CHz-CH3
HS
P
S-C-CH3
II 0
reductose
S-C-CH2-C-CH3
v
-,
Jbr
I
Acetyl/Molonyl Transocylase
-
O-C-CH3
___,
11
II
HO-C-CH C-CoA Acetyl /Malonyl
2 Transocylase
OH
+
2-
CH3- (CH2)14- COOH
Figure 3. Schematic representation of palmitic acid synthesis by chicken liver fatty acid synthase. -SH is 4’-phosphopantetheine, -SH is a cysteine sulfhydryl, and -OHis a serine hydroxyl. The initial sequence of reactions, malonyl transacylase, b-ketoacyl synthase, P-ketoacyl reductase, dehydratase, and enoyl reductase, is repeated seven times to give enzyme-bound palmitic acid.
onto the enzyme as an oxyester on serine and are then passed to 4’-phosphopantetheine, where a thioester linkage is formed; the condensation, reduction, dehydration, reduction (eqs 4-7), and thioesterase reactions occur with the malonyl and unsaturated growing fatty acid chain attached at this site. The saturated growing fatty acid chain is linked covalently to a cysteine sulfhydryl during the loading of the malonyl group and condenses with the malonyl group on 4’-phosphopantetheine. The reaction cycle and location of the bound intermediates are shown in Figure 3. The locations of the serine hydroxyl, 4’-phosphopantetheine, and cysteine on the enzyme are shown in Figures 1 and 2. The evidence supporting this mechanism can be illustrated by studies of the binding of acetyl and malonyl to the enzyme after reaction with acetyl-coA and malonyl-CoA. Early s t ~ d i e s demonstrated ~~-~~ that acetyl binds to both 4’-phosphopantetheine and cysteine, whereas malonyl binds only to the former. If the acetylated enzyme is treated with hydroxylamine, all of the acetyl groups are rapidly removed.29 However, if the acetylated enzyme is denatured and then treated with hydroxylamine, only about 60% of the acetyl groups are removed. Hydroxylamine cleaves thioesters but not normal oxyesters. These results indicate that acetyl is bound to the enzyme as thioesters and an unstable oxyester. Analyses of the binding of a fluorescent analogue of acetyl-coA to the enzyme suggest two different types of sulfhydryl g r o ~ p s . ~ ~These ,~’ (25) Chesterton, C. J.; Butterworth, P. H. W.; Porter, J. W . Arch. Biochem. Biophys. 1968, 126, 864. (26) Phillips, G. T.; Nixon, J. E.; Abramovitz, A. S.; Porter, J. W. Arch. Biochem. Biophys. 1970,138, 357. (27) Phillips, G. T.; Nixon, J. E.; Doreey, J. A,; Butterworth, P. H. W.; Chesterton, C. J.; Porter, J. W . Arch. Biochem. Biophys. 1970, 139, 380. (28) Joshi, V. C.; Plate, C. A.; Wakil, S. J. J. Biol. Chem. 1970, 245, 2857. (29) Cognet, J. A. H.; Hammes, G. G. Biochemistry 1983, 22, 3002.
results are consistent with the mechanism described above. The amino acid sequences of peptides associated with the serine, cysteine, and 4’-phosphopantetheine sites have been ~ b t a i n e d .These ~ sites are indicated in Figure 2. The location of reaction intermediates on the enzyme has been investigated by labeling the growing fatty acid If the chains on the enzyme with radi~activity.~~ thioesterase is cleaved from the enzyme prior to labeling, the labeled intermediates accumulate on the enzyme. After limited trypsinization of the enzyme, enzyme fragments containing either cysteine or 4’phosphopantetheine can be isolated. The intermediates then can be released from the fragments and analyzed by high-performance liquid chromatography. The nature of the intermediates (e.g., extent of saturation, chain length) can be controlled by adjusting the substrate concentrations. For example, if NADPH is present in limited amounts, predominantly 3-hydroxybutyryl-enzyme is formed, and >90% of the 3-hydroxybutyrylis found on 4’-phosphopantetheine. However, when the 3hydroxybutyryl-enzyme is reduced with NADPH, the butyrate is found in approximately equal amounts on cysteine and 4’-phosphopantetheine. The unreduced species, acetoacetate, also is found primarily on 4’phosphopantetheine. This is consistent with the transformation of acetoacetate to butyrate taking place with the intermediate on 4‘-phosphopantetheine and butyrate then being transferred to cysteine. The ratio of saturated intermediate bound on cysteine to that on 4’-phosphopantetheine has been determined as a function of the fatty acid chain length. The ratio is in the range 2-4 for C14or less but drops to 0.3 (30) Cardon, J. W.; Hammes, G. G. Biochemistry 1982, 21, 2863. (31) Cardon, J. W.; Hammes, G. G. J. Biol. Chem. 1983, 258, 4802. (32) Anderson, V. E.; Hammes, G. G. Biochemistry 1985, 24, 2147.
Acc. Chem. Res., Vol. 23, No. 11, 1990 367
Fatty Acid Synthase H
CH3
HH ,LO,
"+,
R
Kinetics of Fatty Acid Synthase
NADPH" pro- 4s
inversion
o=c
4!
pro-S
o=c
0=C
I-
\
pro-R
FH3
HA ,H ," syn e l i m i n a t i o n
NADPH"
(pro-S on C-2)'
pro- 4R
)AH o=c
H X H '
' trans
C H ~
x o=c
pro-S f r o m solvent
Figure 4. Stereochemistry of the reactions catalyzed by chicken liver fatty acid synthase as described in the text.33 pro S from solvent pro S from pro 4s N A D P H
i.ci H-C-H
pro R from pro S on MaiCoA
H-7-H
pro R from pro 4R NADPH
CH,
NADPH, respectively. As an additional benefit, knowledge of the stereochemistry permits synthesis of specifically deuterated or tritiated palmitic acid.
methyl from ACCoA
Figure 5. Origin of the prochiral and methyl hydrogens of fatty acids synthesized by chicken liver fatty acid synthase, as inferred from stereochemical studies of the synthesis of butyrate.33
for cJ6.and is undetectable for C18. (Hydrolysis of palmitic acid does not occur because the thioesterase has been removed from the enzyme.) This suggests that the chain-length specificity of the enzyme is determined by the inability of chain lengths greater than C14 to be transferred to the cysteine. This could be due to steric restrictions on the intermediate binding site associated with the cysteine. The difficulty in chain elongation beyond c16 enhances the probability of termination at this chain length. The mechanism in Figure 3 is consistent with all existing data. However, there is some indication from isotopic distribution studies that the initial acetyl bonding may not be completely ordered, that is, the acetate may react directly with the cysteine and 4'phosphopantetheine, as well as with the serine hydro~yl.~~ The stereochemistry of the sequence of reactions in Figure 3 has been elucidated as shown in Figure 4.33 The condensation of malonyl and acetyl proceeds with inversion of C-2 of malonyl. The next step, the reduction of acetoacetyl-CoA to 3-hydroxybutyryl-CoA, occurs with the transfer of the pro-4S hydrogen of NADPH, a conserved feature of all fatty acid syntheses. The elimination of water is syn with the p r o 3 hydrogen from C-2 being removed. The reduction of crotonylCoA to butyryl-CoA utilizes the pro-4R hydrogen of NADPH which is transferred to the pro-3R position of butyryl-CoA, with the concomitant addition of a solvent proton to the pro-2S position of butyryl-CoA. As shown in Figure 5, with the exception of the methyl hydrogens on the terminal carbon derived from the primer acetyl group, the hydrogens on the carbons of the fatty acid synthesized can be traced to a unique source. The pro-R and pro-S hydrogens on the even carbons are derived from the pro-S hydrogen of malonyl-CoA and from solvent, respectively, whereas the pro-R and pro-S hydrogens on the odd carbons are derived from the pro-4R and pro-4S hydrogens of (33) Anderson, V. E.; Hammes, G . G . Biochemistry 1984, 23, 2088.
The kinetics of fatty acid synthesis are quite complex, as might be expected, and a variety of techniques must be used to obtain information about the elementary steps in the catalytic process. The steady-state kinetics follow a typical Michaelis-Menten rate law for the three substrates, acetyl-coA, malonyl-CoA, and NADPH.34 An unusual feature of the rate law is that malonyl-CoA is a competitive inhibitor of acetyl-coA and acetyl-coA is a competitive inhibitor of malonyl-CoA. The rate law is in accord with the mechanism in Figure 3, with acetyl-coA and malonyl-CoA competing for the same "loading" sites on the enzyme. The steady-state reaction is assayed by following the decrease in absorbance as NADPH is oxidized to NADP'. If phosphotransacetylase is used to rapidly scavenge CoA from the assay mixture, the synthesis of fatty acid stop^.^^^^ Although CoA is a product of the reaction, it is essential for catalysis. In fact, at low concentrations, CoA is an activator, and at high concentrations, it is an inhibitor. The explanation for this unusual result is that CoA rapidly exchanges acetyl and malonyl groups with the enzyme. This permits the rapid exchange of acetyl and malonyl groups between binding sites on the enzyme. Without CoA, for example, acetyl could form a thioester with 4'-phosphopantetheine which would dissociate from the enzyme very slowly, thereby preventing malonyl from forming a thioester with 4'-phosphopantetheine and subsequent synthesis of palmitic acid. If the rates of the catalytic steps are assumed to be independent of the chain length of the fatty acid, the turnover number of the overall reaction, kcat, can be written as l/kcat = ( l / k l
+ 7/k2 + 7/k3 + 7/k4 + 7/k5 + 7/ks + l/k7) (8)
where kl, k2, kg,It4, It5, k6, and k7 are the turnover numbers for acetyl transacylase, malonyl transacylase, @-ketoacylsynthase, @-ketoacylreductase, dehydratase, enoyl reductase, and thioesterase, respectively. The overall reaction has been dissected into elementary steps through the use of fast-reaction techniques (cf. ref 37). The acetylation of fatty acid synthase by acetyl-coA (Ac-CoA)has been studied with quenchedflow methods.29 The kinetic data are consistent with a mechanism involving relatively rapid binding of AcCoA to the enzyme followed by acetylation:
+
Ac-COA E F+E-Ac-COA
k k-i
EAcCOA * EAc + CoA (9)
At pH 7.0 and 23 "C,k, = 43 s-l and = 103 s-l. The removal of acetyl from the enzyme by CoA is quite rapid, consistent with its effect on the steady-state (34) Cox,B. G.; Hammes, G. G . Proc. Natl. Acad. Sci. U.S.A. 1983, 80, 4233. (35) Linn, T. C.; Srere, P. A. J . B i d . Chem. 1980, 255, 10676. (36) Stern, A.; Sedgwick, B.; Smith, S. J. Biol. Chem. 1982,257,799. (37) Hammes, G. G . Enzyme Catalysis and Regulation; Academic Press: New York, 1982; p 220.
368 Acc. Chem. Res., Vol. 23, No. 11, 1990 kinetics. The above reaction characterizes the overall rate of acetylation of the enzyme; however, three different binding sites for acetyl exist on the enzyme. Information about the rate of acetylation of the individual binding sites can be obtained by specifically blocking either the cysteine or cysteine and 4'-phosphopantetheine through chemical modification of the protein.% As expected, formation of the oxyester with serine is relatively rapid, with kl = 150 s-l. If acetylation of the three binding sites is random, then for binding to 4'-phosphopantetheine, k1 < 110 s-l; and for binding to cysteine, kl < 43 s-l. If the reaction is strictly ordered as in Figure 3, the rate constant for intramolecular transfer of acetyl from serine to 4'-phosphopantetheine is 37 s-l. The hydrolysis of the palmitoyl-enzyme has been studied by using palmitoyl-CoA as a substrate (W. Liu and G. G. Hammes, unpublished results). The turnover number is approximately 2 s-l. Whether or not this reaction is a good model for the hydrolysis of palmitoyl-enzyme is unknown. As expected, this is a relatively slow reaction: otherwise the growing fatty acid chain might be terminated prematurely. If the turnover numbers of the individual reactions are assumed to be independent of the fatty acid chain length, the overall turnover number can be estimated (eq 8): k,, < (1/43 + 7/31 + 7/17 + 7/37 + 1/2)-l = 0.7 s-l, with the malonyl transacylase and dehydratase turnover numbers being unknown. The measured which is in steady-state turnover number is 0.8 remarkably good agreement. An assessment of how the turnover numbers of the individual enzymes vary with the fatty acid chain length can be obtained by examining the distribution of chain lengths covalently bound to the enzyme during catalys ~ s As . ~ previously ~ discussed, if the thioesterase is removed from the enzyme complex, covalently bound intermediates accumulate on the enzyme and the nature of the intermediates can be varied by adjusting the substrate concentrations. The distribution of intermediates can be predicted by simple kinetic models. A detailed analysis suggests that the initial condensation and reduction steps are slower than the analogous reactions with longer chain length intermediates. Although a unique solution cannot be obtained, models in which the rate constants increase by about a factor of 2 as the chain length increases by 2 carbons are consistent with the data. For this model, the calculated value of k, is
